Tutoial - Part 2 TutoRial - Part 2

Marine Ecosystem Dynamics - 2025

Pipes

Pipes, expressed as %>% or |>, are very useful and make our code clearer. Using pipes, our data flow from one function to another.

Exercises

• Rewrite these chunks of code using the pipes

sum(c(2.2,4.1,2,pi))

Solution

c(2.2,4.1,2,pi) |> sum()
# OR
c(2.2,4.1,2,pi) %>% sum()

round(sum(c(2.2,4.1,2,pi)))

Solution

c(2.2,4.1,2,pi) |> sum() |> round()
# OR
c(2.2,4.1,2,pi) %>% sum() %>% round()

round(sum(c(2.2,4.1,2,pi)), digits = 3)

Solution

c(2.2,4.1,2,pi) |> sum() |> round(digits = 3)
# OR
c(2.2,4.1,2,pi) %>% sum() %>% round(digits = 3)

Tidy the data with tidyr

As seen in the slides, a tidy table has:

1. Each variable in its own column
2. Each observation in its own row

To reach this, tidyr has 4 key functions:

1. pivot_longer
2. pivot_wider
3. unite
4. separate

Exercises

• If this is not done yet, download the dataset zooplankton_seasonality.csv

• Import the dataset in your environment

• Is this dataset a tidy dataset?

Solution

First 6 rows of the dataset zooplankton_seasonality

Month_abb	Year	Station	Coordinates	Group	Taxa	Biomass
Jan	2009	BY15	20.05000/57.33333	Copepoda	Acartia	6.650319
Jan	2009	BY31	18.23333/58.58812	Copepoda	Acartia	1.816994
Jan	2009	BY5	15.98333/55.25000	Copepoda	Acartia	5.562097
Jan	2009	BY15	20.05000/57.33333	Copepoda	Centropages	5.738562
Jan	2009	BY31	18.23333/58.58812	Copepoda	Centropages	1.228759
Jan	2009	BY5	15.98333/55.25000	Copepoda	Centropages	14.405224

Each variable has its own column
Each variable has its own row
Coordinates has 2 values

• Separate the column Coordinates in 2 news columns: Longitude and Latitude

Solution

library(tidyr)
zooplankton |>
  separate(Coordinates, into = c("Longitude", "Latitude"), sep = "/")

First 6 rows of the transformed dataset zooplankton_seasonality

Month_abb	Year	Station	Longitude	Latitude	Group	Taxa	Biomass
Jan	2009	BY15	20.05000	57.33333	Copepoda	Acartia	6.650319
Jan	2009	BY31	18.23333	58.58812	Copepoda	Acartia	1.816994
Jan	2009	BY5	15.98333	55.25000	Copepoda	Acartia	5.562097
Jan	2009	BY15	20.05000	57.33333	Copepoda	Centropages	5.738562
Jan	2009	BY31	18.23333	58.58812	Copepoda	Centropages	1.228759
Jan	2009	BY5	15.98333	55.25000	Copepoda	Centropages	14.405224

• Combine the column Groupand Taxa into a new column Group_Taxa and save the dataframe as tidy_df

Solution

library(tidyr)
tidy_df <-
  zooplankton |>
  separate(Coordinates, into = c("Longitude", "Latitude"), sep = "/") |> 
  unite("Group_Taxa", c(Group, Taxa))

First 6 rows of the transformed dataset zooplankton_seasonality

Month_abb	Year	Station	Longitude	Latitude	Group_Taxa	Biomass
Jan	2009	BY15	20.05000	57.33333	Copepoda_Acartia	6.650319
Jan	2009	BY31	18.23333	58.58812	Copepoda_Acartia	1.816994
Jan	2009	BY5	15.98333	55.25000	Copepoda_Acartia	5.562097
Jan	2009	BY15	20.05000	57.33333	Copepoda_Centropages	5.738562
Jan	2009	BY31	18.23333	58.58812	Copepoda_Centropages	1.228759
Jan	2009	BY5	15.98333	55.25000	Copepoda_Centropages	14.405224

• Create a wide table with columns having the Biomass values for each Group_Taxa and save the dataframe as wide_df

Solution

library(tidyr)
wide_df <-
  tidy_df |> 
  pivot_wider(names_from = Group_Taxa, values_from = Biomass) 

First 6 rows of the transformed dataset zooplankton_seasonality

Month_abb	Year	Station	Longitude	Latitude	Copepoda_Acartia	Copepoda_Centropages	Copepoda_Pseudocalanus	Copepoda_Temora	Rotatoria_Synchaeta	Copepoda_Eurytemora	Rotatoria_Keratella	Cladocera_Bosmina	Cladocera_Evadne	Cladocera_Podon
Jan	2009	BY15	20.05000	57.33333	6.650319	5.7385615	10.522882	9.725488	0.3921570	NA	NA	NA	NA	NA
Jan	2009	BY31	18.23333	58.58812	1.816994	1.2287586	5.633984	4.993465	0.4705890	NA	NA	NA	NA	NA
Jan	2009	BY5	15.98333	55.25000	5.562097	14.4052240	21.594775	45.738529	0.3921570	NA	NA	NA	NA	NA
Jan	2010	BY15	20.05000	57.33333	2.467319	0.3071893	13.601301	7.549021	0.1568628	NA	NA	NA	NA	NA
Jan	2010	BY31	18.23333	58.58812	2.248367	0.3856208	2.660128	8.418301	0.4117650	0.0849674	NA	NA	NA	NA
Jan	2011	BY15	20.05000	57.33333	5.065367	2.9803908	49.660135	36.431384	0.5490210	NA	NA	NA	NA	NA

Data handling with dplyr

After finishing tidying the data, we often use the dplyr package to process our data.

Exercises

• What is the class of the Year columns of the tidy_df dataframe?
  If they are not numeric, mutate them as numeric values.

Solution

str(tidy_df)
#> tibble [2,956 × 7] (S3: tbl_df/tbl/data.frame)
#>  $ Month_abb : chr [1:2956] "Jan" "Jan" "Jan" "Jan" ...
#>  $ Year      : chr [1:2956] "2009" "2009" "2009" "2009" ...
#>  $ Station   : chr [1:2956] "BY15" "BY31" "BY5" "BY15" ...
#>  $ Longitude : chr [1:2956] "20.05000" "18.23333" "15.98333" "20.05000" ...
#>  $ Latitude  : chr [1:2956] "57.33333" "58.58812" "55.25000" "57.33333" ...
#>  $ Group_Taxa: chr [1:2956] "Copepoda_Acartia" "Copepoda_Acartia" "Copepoda_Acartia" "Copepoda_Centropages" ...
#>  $ Biomass   : num [1:2956] 6.65 1.82 5.56 5.74 1.23 ...

Longitude and Latitude are characters

library(dplyr)
tidy_df |> 
  mutate(Year = as.numeric(Year))

• Then, kepp all Year between 2012 and 2015

Solution

library(dplyr)
tidy_df |> 
  mutate(Longitude = as.numeric(Longitude),
         Latitude = as.numeric(Latitude)) |> 
  filter(Year %in% 2012:2015)

• Then, only keep the data from the Station BY31

Solution

library(dplyr)
tidy_df |> 
  mutate(Longitude = as.numeric(Longitude),
         Latitude = as.numeric(Latitude)) |> 
  filter(Year %in% 2012:2015) |> 
  filter(Station == "BY31")

# OR

tidy_df |> 
  mutate(Longitude = as.numeric(Longitude),
         Latitude = as.numeric(Latitude)) |> 
  filter(Year %in% 2012:2015,
         Station == "BY31")

• Then, select all columns except Longitude and Latitude

Solution

library(dplyr)
tidy_df |> 
  mutate(Longitude = as.numeric(Longitude),
         Latitude = as.numeric(Latitude)) |> 
  filter(Year %in% 2012:2015,
         Station == "BY31") |> 
  select(-Longitude,
         -Latitude)

• Then, rename Month_abb as Month

Solution

library(dplyr)
tidy_df |> 
  mutate(Longitude = as.numeric(Longitude),
         Latitude = as.numeric(Latitude)) |> 
  filter(Year %in% 2012:2015,
         Station == "BY31") |> 
  select(-Longitude,
         -Latitude) |> 
  rename(Month = Month_abb)

• Then, group_by: Month and Group_Taxa and take the Biomass average and standard deviation and save the dataframe as summarized_df

Solution

library(dplyr)
summarised_df <-
  tidy_df |> 
  mutate(Longitude = as.numeric(Longitude),
         Latitude = as.numeric(Latitude)) |> 
  filter(Year %in% 2012:2015,
         Station == "BY31") |> 
  select(-Longitude,
         -Latitude) |> 
  rename(Month = Month_abb) |> 
  group_by(Month, Group_Taxa) |> 
  summarise(average = mean(Biomass),
            standard_deviation = sd(Biomass))

Ploting the data with ggplot2

In this part, we will build a plot step by step using the grammar of graphic in ggplot2

• Load the package and only keep the values for the copepod Acartia from the summarised_df dataset in a new dataset called acartia

Solution

library(ggplot2)
acartia <-
  summarised_df |> 
  filter(Group_Taxa == "Copepoda_Acartia")

• Initiate a ggplot with the dataset acartia with the Month as the x-axis and the average biomass as the y-axis

Solution

ggplot(data = acartia,
       mapping = aes(x = Month, y = average))

• Add a barplot geometry to the plot

Solution

ggplot(data = acartia,
       mapping = aes(x = Month, y = average)) +
  geom_bar(stat = "identity") # <- this is needed for the barplot geometry

• Arrange the bar from the lowest to the highest values

Solution

ggplot(data = acartia,
       mapping = aes(x = reorder(Month, average), y = average)) +
  geom_bar(stat = "identity") # <- this is needed for the barplot geometry

• Add a color filling in the bars according the Month

Solution

ggplot(data = acartia,
       mapping = aes(x = reorder(Month, average), y = average, fill = Month)) +
  geom_bar(stat = "identity") # <- this is needed for the barplot geometry

• Change the axis as Biomass and Month, and add a title

Solution

ggplot(data = acartia,
       mapping = aes(x = reorder(Month, average), y = average, fill = Month)) +
  geom_bar(stat = "identity") + # <- this is needed for the barplot geometry
  labs(x = "Month", y = "Biomass", title = "My nice ggplot")